Helm

And then we just fill whatever that is needed.

So that goes for Helm charts, that goes for Terraform files, that goes for CICD configurations.

So those are kind of like I have this thing, but I want to multiply it.

So it just follows the pattern, essentially.

And it makes several good arguments that I'm not against and actually agree with some of these for various reasons.

Chief among those is that managing lots of little services through Terraform and just through GitOps with the available tooling is not as convenient as deploying architectures via Helm and Argo CD and Flux, for example.

Then there are platforms on top of Kubernetes, of course, that I think are critical.

Crossplane is frequently mentioned, but I think that there's a very mature operator ecosystem that means you don't have to go all in on Crossplane, for example.

Ship smarter, not harder.

Affordable Kubernetes C CD with GitHub Actions plus Helm This guide is for freelancers and small teams who need reliable, inexpensive delivery to Kubernetes.

96.

Effortlessly deploy a Docker image to ECR with GitHub Actions in this post.

So it's basically our, uh, Thales flight edge, uh, distro, if you want to see it this way.

Um, but it's, uh, it's really trying to, um, to keep very standard using a technology such as, uh, Helm to package, uh, the software, for example, in, on top of the containers.

Um, and then as Fred was saying, we have a set of managed services.

Uh, we make these available in this, uh, in this cloud that we offer to the, to the airlines to deploy their own services and also that we use to deploy our services.

Because the interviewer will ask, "How do the Vault agents actually get into your pod?"

Right, and the trap here is saying, "Oh, I just manually add the Vault sidecar specs into my Helm chart."

If you're manually copy-pasting sidecars across five hundred microservices, that's a maintenance nightmare.

What if Vault needs a security patch?

Modern Istio admission webhooks catch this, but historically it caused huge outages.

Terrifying.

Let's move to section five: packaging and deployments, Helm.

What's the difference between chart.version and appVersion?

Yeah, three updates.

Um, first is, uh, as you know, we added support for Helm, native Helm support, uh, to Atmos.

Uh, with native Helm support in Atmos, it can automatically deploy directly to Kubernetes, or it can deploy to a target.

What is a target? Well, we s-support configuring, uh, Git targets, so it can render man- the Helm chart to a manifest, uh, that gets committed to a, a Git repository.

For example, if you're using a Git repo for, um, uh, for Argo CD or Flux.

Now, if you're doing direct deployments, direct CD, as we call it, to the cluster, one of the nice things to be able to do is to wait for failure.

So with this change, now Atmos will, uh, immediately probe and find out the last, uh, couple log messages from, uh, the crash looping pod and display those as part of the exit message in Atmos, something like this.

So it just makes it easier both for developers and their agents to quickly triage what's going on.

Today's guest is Corio Daniel.

Corey is the CEO and co-founder of MassDriver, an internal developer platform and platform orchestrator that helps teams turn Terraform, OpenTofu, and Helm into governed self-service infra.

A software engineer and cloud ops veteran, he has spent more than 20 years building teams and startups working with cloud infra.

In this episode, we also discuss thinking about identities and permissions for agents, testing cloud infra rollouts in an agentic loop, and how we're all citizen developers in one way or another.